Terahertz photoconductive antennas have been used for more than two decades. Since its invention in 1984, minor modifications have been made to the antenna structures; however, details of the antenna design and the parallel micro-strip-line electrodes, which form the basic electrode structure of the conventional photoconductive antenna, have not been modified and are still being used.
FIG. 1 is a prior art terahertz photoconductive antenna structure 100. The parallel electrodes 105 are typically fabricated by depositing gold layers into two parallel trenches. Although the trench depth in FIG. 1 and the gold electrode thickness are labeled as 650 nm and 520 nm, the depth and thickness of the commercial photoconductive antennas have not been optimized, and there is no standard for these values. Therefore, it is not uncommon to see a large variation in these parameters from photoconductive-antenna manufacturers, and there are few guidelines for the fabrication of electrodes. The gold electrodes can often be excessively deposited, so that their thickness exceeds 1 μm. In commercial photoconductive antennas, the gold layers (electrodes) directly contact the trench walls, so that electric currents can flow from the electrode through the sidewall of the trench.
FIG. 2A is a prior art diagram representing the generation of a terahertz pulse using a femto-second laser and a photoconductive antenna. FIG. 2B is a prior art diagram representing the positive and negative charges between the electrodes during the generation of a terahertz pulse. FIG. 3 is a prior art diagram of a cross-sectional view of photoconductive antenna showing the photocurrents, bias currents and thermal currents during the generation of a terahertz pulse. FIGS. 2A and 3 illustrate that the terahertz pulse can be produced by illuminating a semiconductor slab (e.g., a GaAs substrate) with a femto-second laser beam. The laser pulse can generate a surface plasma, consisting of positive charges and negative charges. This oscillating surface plasma is known as a surface plasmon. The oscillating positive and negative charges can generate the terahertz pulse.
If the positive and negative charges recombine immediately after they are produced, the intensity of the terahertz pulse becomes very weak. Therefore, in order to minimize the charge recombination, a bias voltage can be applied to the electrodes, which can create an electric field that separates the positive charges from the negative charges (see FIG. 2(B1)). The positive charges will be attracted to the negative electrode, and the negative charges will be attracted to the positive electrode. However, when these charges arrive at the electrodes, they will ordinarily be discharged. To prevent such a discharge, the polarity of the electrodes can be switched right before the charges touch the electrodes, or just before they collide and recombine (see FIG. 2(B2) and FIG. 2(B3)). In other words, an AC bias voltage with an optimum frequency can substantially enhance the oscillation amplitude of the plasmon (the photocurrent) so that it increases the terahertz pulse strength.
The ac bias voltage, however, can result in substantial bias current flowing between the electrodes. This bias current, along with the photocurrent, can generate considerable Joule heating. The Joule heating, together with the thermal energy provided by the femto-second laser beam, can create thermal electric currents, which are incoherent in nature. The photocurrent, bias currents and thermal electric currents can all produce Joule heat, and the Joule heat can create more thermal currents, which can produce blackbody radiation, such as incoherent terahertz beams and infrared beams.
The thermal electric currents can also disrupt the coherency of the photocurrent and the bias currents, so that it reduces the strength of the coherent terahertz beam, and enhances the incoherent terahertz beam. This further increases the heating and the thermal electric currents. FIG. 4 is a prior art schematic diagram the explains how the photocurrent, bias-current, and thermal currents affect the production of a coherent terahertz beam and an incoherent terahertz beam. Specifically, FIG. 4 depicts the complex recursive process and the interactions among the three different currents. If the heat produced through this recursive process is excessive, it will eventually destroy the photoconductive antenna.
When a thermal electron (or electron wave function) travels perpendicular to the electrodes, in between a pair of parallel electrodes, the electron (or electron wave function) is likely to follow a bouncing ball trajectory or a standing wave pattern. FIG. 5A represents a bouncing ball trajectory for an electron wave function. FIG. 5B represents a standing wave pattern for an electron wave function. Therefore, the particle (the electron) or wave (the electron wave) is likely to be trapped in between the electrodes, unless the particle or wave travels at an oblique angle, such as in FIG. 6A and FIG. 7A. FIG. 6A represents a traveling ball mode trajectory for an electron wave function, and, similarly, FIG. 7A represents a non-chaotic trajectory, such as the traveling ball mode, for an electron wave function. Therefore, with the parallel electrode geometry, a large number of thermal electrons that flow incoherently can be trapped in between the electrodes and disrupt the photocurrent. FIG. 8A represents a slowly traveling, or virtually trapped, wave pattern for a traveling ball mode trajectory for an electron wave function.
Additionally, FIG. 6B represents a trapped ball mode trajectory for an electron wave function, and the associated FIG. 8B represents a standing wave mode for a trapped ball mode trajectory for an electron wave function. Similarly to FIG. 6A and associated FIG. 8A, FIG. 6B and FIG. 8B show how a large number of thermal electrons that flow incoherently can be trapped in between the electrodes and disrupt the photocurrent, and may not allow traveling wave pattern.
Consequently, a conventional photoconductive antenna with a pair of parallel electrodes is highly inefficient in converting the femto-second laser pulse into a terahertz beam; and, therefore, is not efficient in producing a strong, coherent terahertz beam. Instead, because of the above-mentioned problems, the antenna structure with the conventional electrodes predominantly produces incoherent terahertz beams, and the efficiency of the conventional photoconductive antenna is therefore very poor.
In summary, the conventional terahertz photoconductive antennas have the following limitations and disadvantages: (1) the design parameters, such as the trench depth, the thickness of the electrode, and the thickness of the substrate, are not optimized; (2) the conventional photoconductive antenna with a pair of parallel electrodes produces a very weak, coherent terahertz beam (<<1 mW); (3) with a strong pump-laser beam and a large bias voltage applied to the electrodes, they produce excessively incoherent terahertz beams, which lead to the destruction of the photoconductive antenna; and (4) as a result, the lifetime of the conventional photoconductive antenna is short.
Accordingly, there remains a need in the art for an improved design of a photoconductive antenna that can dramatically improve its conversion efficiency, and produce a strong, coherent terahertz beam.